UniECG: Understanding and Generating ECG in One Unified Model

Haoyu Wang; Hongyan Li; Jiarui Jin; Jun Li; Shenda Hong; Xiang Lan

REVIEW 3 major objections 4 minor 2 cited by

UniECG: Understanding and Generating ECG in One Unified Model

T0 review · 3 major / 4 minor · reviewed 2026-08-04 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read UniECG proposes the first unified ECG model that can both produce evidence-based interpretations of ECG signals and images, and generate signal-level ECGs from text descriptions.

desk verdict Reasonable two-stage idea, but the generation claim rests on one qualitative example and a distribution shift the paper never addresses. read the letter →

arxiv 2509.18588 v2 pith:HFG7GHFG submitted 2025-09-23 cs.CL

Jiarui Jin , Haoyu Wang , Xiang Lan , Jun Li , Hongyan Li , Shenda Hong This is my paper

classification cs.CL

keywords ECGmultimodallargelanguagemodeltext-to-ECGgenerationinterpretationlatentspacealignmentdiffusiontwo-stagetrainingmedicaleducation

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper introduces UniECG, a single multimodal large language model that attempts to do both directions of the ECG task: given an ECG signal or image, it produces an evidence-based diagnostic explanation, and given a textual description, it generates a signal-level 12-lead ECG waveform. The proposal is a two-stage training recipe: first the model is fully fine-tuned to interpret ECG data, then its newly added [ECG] tokens are aligned in latent space with the text encoder of a frozen diffusion-based ECG generator (DiffuSETS), so the same LLM can emit tokens that the diffusion decoder turns into waveforms. The paper reports that understanding performance stays in the same range as the current grounding-focused model while gaining generative ability absent from previous ECG multimodal models. The motivation is interactive ECG education and case-based learning, not clinical diagnosis.

What carries the argument

The load-bearing mechanism is latent-space alignment: the LLM vocabulary gains n special [ECG] tokens, and after the model is frozen, only these token embeddings are fine-tuned (Eq. 2). At inference, the final-layer hidden states of the [ECG] sequence are passed through an encoder–decoder Transformer mapper with learnable queries, and mean squared error (Eq. 3) pulls its outputs into the embedding space of DiffuSETS's frozen text encoder. DiffuSETS — a diffusion model that generates 12-lead ECGs conditioned on text reports — then decodes these aligned embeddings into waveforms. The trick is that the second training stage needs no ECG data at all, only MIMIC-IV-ECG text reports converted to '

What would settle it

Generate ECGs from a set of textual reports with known reference classes (e.g., atrial fibrillation, left bundle branch block, heart rates 60 vs 120 bpm), then run a pretrained ECG classifier on the generated signals and compare its label agreement and morphology statistics (QRS duration, RR irregularity) against real MIMIC-IV-ECG waveforms with the same text conditions. If the classifier cannot distinguish conditions or morphology metrics deviate substantially from real data, the alignment has not transferred to the diffusion decoder.

Watch

Extended reading notes

Core claim

The central claim is that one model can be made capable of both evidence-based ECG interpretation and text-conditioned ECG signal generation, without sacrificing either. The two-stage design first teaches ECG-to-text reasoning by full-parameter fine-tuning on the ECG-Grounding dataset, then freezes the LLM and teaches text-to-ECG by adding trainable [ECG] tokens whose hidden states are projected by a lightweight mapper toward DiffuSETS's frozen text-encoder embeddings; the pretrained diffusion decoder then synthesizes the signal. The paper shows quantitative understanding scores (e.g., overall average 63.98 vs 69.59 for the current best comparison) and qualitative generation examples, arguin

Load-bearing premise

The generation half of the central claim rests on the assumption that aligning [ECG] token hidden states with DiffuSETS's frozen text-encoder embeddings (using only text, no waveforms) yields embedding-space regions from which the diffusion decoder produces condition-faithful, clinically plausible ECG signals; the paper presents only qualitative support for that transfer.

Editorial extensions

If this is right

A single model can switch between explaining a real ECG and generating a synthetic one based solely on user wording, enabling dialogue-style case-based learning.
Signal-level (not image-only) outputs mean generated ECGs can be fed to downstream deep-learning models, potentially supporting data augmentation for rare or under-represented cardiac conditions.
The generation skill is acquired without waveform supervision, since the frozen diffusion decoder supplies signal fidelity while the LLM learns only when and how to emit [ECG] tokens.
The decoupling means understanding fine-tuning and generation alignment can be done independently; one can upgrade the generator without retraining interpretation.
If the approach holds, ECG education tools could let a student request 'an ECG with atrial fibrillation and a heart rate of 110' and then question the model about the trace it just produced.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

An obvious next step the paper does not perform is quantitative evaluation of generations (e.g., classifier agreement with the conditioning text, beat-level morphology metrics); the current evidence for 'high fidelity' is qualitative.
The same two-stage latent-alignment recipe could transfer to other physiological time series where a text-conditioned diffusion model exists, such as PPG or EEG, turning a multimodal LLM into a signal synthesizer without waveform training.
Because the generation stage is trained only on reports, the model's generative behavior is bounded by the diversity of MIMIC-IV-ECG text; novel conditions not covered by report phrasing may not be synthesizable.
If the alignment lands in low-density regions of the frozen embedding space, generations could look plausible yet be clinically inconsistent — a failure mode a user-facing educational tool would need to guard against.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this is a neat but under-evidenced idea. The paper combines ECG explanation and generation in one LLM by adding [ECG] tokens and learning a mapper from their hidden states to the text-embedding space of the frozen DiffuSETS diffusion model. As far as I know, no prior ECG model does both in one pass, so the setup is a real contribution. The two-stage training is clearly described, and freezing the LLM while only adding tokens in stage 2 is sensible. Understanding results are moderately positive: they beat PULSE by a wide margin and match GEM on lead coverage, though they trail on most other dimensions.

The soft spots are real and load-bearing on the generation side. There is no quantitative evaluation of generated ECGs: no fidelity, diversity, condition-match, or clinical correctness metrics, and no comparison to DiffuSETS itself, despite it being the same group's prior work. A single qualitative example with atrial fibrillation is not enough to support "high-fidelity" or "remarkable generation ability."

The stress-test concern is valid. Eq. (3) aligns mapper outputs to the text embedding of the full MIMIC-IV-ECG report after the LLM has already seen that full report in context. At inference, users give short prompts like "atrial fibrillation," so the model generates a short prefix and then emits [ECG] tokens conditioned on that prefix, which is outside the training distribution. The paper does not address this exposure-bias/distribution-shift problem. If the mapper generalizes poorly, the generated ECG will not match the condition. This is testable and should be checked with varying prompt lengths.

The understanding evaluation uses DeepSeek V3.1 as an unvalidated judge, with no error bars or human agreement. That weakens the quantitative claims, though the relative improvements over PULSE are large enough to be plausible.

Reproducibility is also a problem: no code or checkpoints are available, and the base LLM and key hyperparameters are unspecified. That makes verification harder.

To the paper's credit, it is transparent about being an educational aid, not a clinical tool, and it acknowledges notable limitations in the conclusion. The citation pattern leans heavily on the same group's prior work, but that is not itself a flaw—it just makes the missing comparison to DiffuSETS more conspicuous.

I would send this to peer review rather than desk reject. The unified-model idea is worth exploring, and a referee can push for generation metrics and an analysis of short-prompt generalization. As submitted, the central generation claim is unsupported; the paper reads like a promising technical note, not a finished system.

Referee Report

3 major / 4 minor

Summary. UniECG is a two-stage ECG multimodal large language model. In stage 1, the LLM is fine-tuned on ECG signal-image-text triplets to produce evidence-based interpretations. In stage 2, the LLM is frozen, special [ECG] tokens are added, and a lightweight mapper is trained to align the hidden states of those tokens with the text-encoder embeddings of DiffuSETS, a pretrained text-conditioned ECG diffusion model. At inference, the model chooses between interpretation and generation based on the user input; generation is performed by feeding the mapped [ECG] token states into the frozen DiffuSETS decoder. The paper reports quantitative understanding results on the ECG-Grounding benchmark scored by DeepSeek V3.1, and qualitative generation results in a single dialogue example.

Significance. If the generation capability were quantitatively supported, a single model that both interprets real ECGs and synthesizes signal-level ECG waveforms from textual prompts would be a meaningful step beyond prior ECG MLLMs, which are understanding-only. The two-stage design—decoupling understanding from generation via latent-space alignment to a frozen diffusion decoder—is an elegant way to add generation without catastrophic forgetting, and it builds on prior work (GILL, DiffuSETS) in a sensible way. The authors explicitly position the model as an educational aid rather than a clinical diagnostic system, which is appropriate. However, the paper's central novelty is the unified model, and the evidence for the generation half is currently anecdotal. The understanding evaluation also needs more rigor before the trade-off against GEM can be interpreted. Strengths include the clear two-stage formulation, the use of publicly available datasets, and the promise of released code and checkpoints.

major comments (3)

[§2.2, Eq. (3)] The alignment loss is trained on full MIMIC-IV-ECG reports: the query is 'Generate an ECG of ...' and the answer is 'This is an ECG of ...' with the complete report filling the ellipsis. The [ECG] hidden states are therefore computed after the LLM has already seen the full report in context. At inference, users provide short prompts (e.g., 'atrial fibrillation'), and the LLM must generate its own answer prefix autoregressively before emitting [ECG] tokens. This is a distribution shift and an exposure-bias problem that is not addressed. The claim that minimizing MSE between mapped [ECG] states and DiffuSETS text embeddings transfers to short, self-generated prompts is an untested assumption. Please provide quantitative generation results on both full-report conditions and short-prompt conditions, including condition-alignment metrics, signal fidelity, and clinical correctness.
[§3.2, Tables 1–2] The understanding results are point estimates from a single scoring run by DeepSeek V3.1, with no validation of the LLM judge, no error bars, and no repeated sampling. UniECG is below GEM on most metrics, including Diagnosis Accuracy (66.80 vs. 72.30), ECG Feature Grounding (69.44 vs. 73.78), and Average (63.98 vs. 69.59). The statement that UniECG 'largely retains' its interpretation ability is therefore not quantitatively supported; at best the tables show a trade-off. Report mean±std over multiple runs or provide a statistical test, and include a human-annotation or independent-judge validation for the metrics.
[§3.2, Fig. 2] The paper's headline claim includes 'high-fidelity ECG signals with fine-grained temporal features', but the only support is a single qualitative figure. There is no quantitative evaluation of generation: no signal fidelity metrics (e.g., morphology error, SNR, Fréchet distance), no condition-alignment metrics, no diversity measure, and no comparison to the underlying DiffuSETS baseline. Since generation is half of the claimed contribution and is the primary novelty over prior ECG MLLMs, this is a load-bearing omission. Please add a generation benchmarking section with quantitative metrics and a comparison to DiffuSETS (and ideally to the training data distribution).

minor comments (4)

[§2.2, Eq. (2)] The notation in Eq. (2) conditions on s_{1:t}, but the stage-2 training is described as text-only and no signal s appears in the surrounding text. The conditioning variable should be the text tokens of the query/answer prefix; please correct.
[§3.1] The ECG-Grounding benchmark is scored by DeepSeek V3.1; please specify whether this judge was validated against human ratings, and report the prompt used for scoring.
[§1] The introduction mentions 'GPT-5' as an example of a unified model, but this model is not established in the literature as of the submission date. Please cite concrete released models (e.g., GPT-4V, LLaVA, BLIP-3) or remove the reference.
[Abstract/§4] The abstract says the model is 'a step toward interactive ECG education', but the paper contains no educational-efficacy evaluation. Please soften the educational claim or add a relevant user study.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found; the generation pipeline is an explicit adapter over a cited pretrained model, and the claimed reductions are not present.

full rationale

UniECG's understanding stage fine-tunes on ECG-Grounding, a dataset co-developed by the authors, but this is standard supervised training with external MIMIC-IV-ECG data, and results are compared against reference systems; no equation reduces the understanding output to the training labels. The generation stage is not a derivation-level tautology: Eq. (3) trains a lightweight mapper to reproduce the frozen DiffuSETS text-encoder embedding Tψ(y), and the ECG waveform is subsequently produced by the externally pretrained DiffuSETS decoder. This is an explicit adapter over a cited generator, not a claim that the LLM itself synthesizes waveforms. The lack of quantitative generation evaluation, and the train/inference mismatch between full-report answers and short user prompts, are real validation gaps that bear on correctness/robustness, but they do not make the derivation circular. Self-citations to DiffuSETS and GEM are present, but the cited components are independent prior artifacts with their own training and falsifiability; no load-bearing conclusion rests solely on an unverified self-citation. Therefore no circular step can be exhibited.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim leans on two same-lab artifacts: the ECG-Grounding benchmark from GEM (Ref 16) and the DiffuSETS generator (Ref 17), both sharing authors with UniECG. Additionally, the alignment hypothesis in Eq. 3 is untested quantitatively, and core hyperparameters (n, L, base LLM identity) are undisclosed.

free parameters (3)

Number of [ECG] tokens n = not reported
Chosen by hand; controls how much generation information the LLM can emit; affects quality of alignment.
Number of mapper query embeddings L = not reported
Chosen by hand; controls capacity of the encoder-decoder Transformer mapper in Eq. 3.
Alignment loss weight lambda = 1
Set to 1 in Eq. 4 without sensitivity analysis; central to balancing token loss and embedding loss.

assumptions (4)

domain assumption ECG-Grounding annotations are correct evidence-based ground truths for ECG interpretation.
Used in Stage 1 to fine-tune the LLM (Section 2.1); if the heartbeat-level labels and diagnostic dialogues contain errors, understanding quality is inherited from those errors.
domain assumption DiffuSETS text encoder and decoder produce high-fidelity, text-consistent ECGs when conditioned on embeddings in its input space.
UniECG delegates all waveform synthesis to frozen DiffuSETS (Sections 2.2 and 2.3); the paper provides no quantitative validation of this premise.
ad hoc to paper Minimizing MSE between [ECG] token hidden states and DiffuSETS text embeddings (Eq. 3) is sufficient to transfer semantic control to the decoder.
This is the core proposed mechanism; only a single qualitative AF example is shown as evidence.
domain assumption DeepSeek V3.1's automated scores reflect clinical quality of ECG explanations.
All understanding metrics in Tables 1 and 2 are scored by DeepSeek V3.1 without human validation or comparison to expert labels.

how reviews work

0 comments

Cite this review

Pith. "Pith review of UniECG: Understanding and Generating ECG in One Unified Model." pith.science (2026). https://pith.science/paper/HFG7GHFG

@misc{pith2026250918588,
  author       = {Pith},
  title        = {Pith review of: UniECG: Understanding and Generating ECG in One Unified Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HFG7GHFG}},
  note         = {Machine review of arXiv:2509.18588}
}

read the original abstract

Electrocardiogram (ECG) interpretation is a fundamental skill in medical education, yet students often need more than static examples to connect waveform evidence with diagnostic reasoning. This paper presents UniECG as a step toward interactive ECG education. UniECG supports two complementary learning interactions: given an ECG signal or image, it generates an evidence-based explanation; given a textual learning objective, it generates a corresponding ECG signal example for case-based learning. The model follows a two-stage design. First, it learns grounded ECG explanation from ECG signal--image--text data. Second, it introduces special ECG generation tokens and aligns their hidden representations with a pretrained text-conditioned ECG diffusion model, enabling controllable signal-level ECG generation. We evaluate UniECG through grounded ECG explanation and generation-oriented qualitative analysis, examining its potential to support explanation and case-based learning. UniECG is intended as an educational aid and a research step toward interactive AI-assisted ECG learning, rather than a clinically validated diagnostic system.

Discussion (0). Sign in to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

R1-SyntheticVL: Is Synthetic Data from Generative Models Ready for Multimodal Large Language Model?
cs.LG 2026-02 conditional novelty 6.0 of 10

A collective adversarial data-synthesis pipeline produces 20K synthetic multimodal training examples whose GRPO-trained 7B model beats several listed open-source MLLMs on reasoning benchmarks.
Reasoning Before Diagnosis: Physician-Inspired Structured Thinking for ECG Classification
cs.AI 2026-05 unverdicted novelty 5.0 of 10

CardioThink applies structured clinical reasoning stages and Structured Set Policy Optimization (SSPO) to ECG classification, yielding higher diagnostic accuracy and more interpretable rationales than direct predictio...

Reference graph

Works this paper leans on

22 extracted references · 8 linked inside Pith · cited by 2 Pith papers

[1]

INTRODUCTION The electrocardiogram (ECG) has become a crucial tool in the clinical diagnosis of cardiac diseases, owing to its non- invasive nature, low cost, and ease of acquisition [1, 2, 3]. While deep learning methods have shown promising results in tasks such as cardiac anomaly detection [4, 5, 6] and rare ECG synthesis [7, 8, 9], these methods are t...

arXiv 2025
[2]

Generate an ECG of

METHODOLOGY 2.1. Learning to Understand ECG Given an ECG signals, an ECG imagex, and its textual captiony, our goal is to fully fine-tune a large language model (LLM) such that it can jointly reason over ECG data and textual inputs, thereby generating clinically meaning- ful responses grounded in ECG interpretation. To achieve this, we leverage the ECG-Gr...
[3]

Datasets We primarily employ the ECG-Grounding and MIMIC-IV- ECG datasets for training and validation in UniECG under- standing and ECG generation tasks

EXPERIMENT AND ANALYSIS 3.1. Datasets We primarily employ the ECG-Grounding and MIMIC-IV- ECG datasets for training and validation in UniECG under- standing and ECG generation tasks. The descriptions of these two datasets are provided as follows: MIMIC-IV-ECG.This publicly available dataset [18] provides a comprehensive collection of 800,035 multiple- lea...
[4]

CONCLUSION In this paper, we introduce UniECG, the first unified model for ECG that is capable of performing both evidence-based ECG understanding and text-guided ECG generation. We propose a two-stage training strategy that equips existing ECG MLLMs with both interpretive and generative capabil- ities, thereby effectively expanding the capability frontie...
[5]

Opportunities and challenges of deep learning methods for electrocardiogram data: A system- atic review,

Shenda Hong, Yuxi Zhou, Junyuan Shang, Cao Xiao, and Jimeng Sun, “Opportunities and challenges of deep learning methods for electrocardiogram data: A system- atic review,”Computers in Biology and Medicine, vol. 122, pp. 103801, July 2020

2020
[6]

Deep learning in ECG diagnosis: A review,

Xinwen Liu, Huan Wang, Zongjin Li, and Lang Qin, “Deep learning in ECG diagnosis: A review,” Knowledge-Based Systems, vol. 227, pp. 107187, Sept. 2021

2021
[7]

Screening for cardiac contrac- tile dysfunction using an artificial intelligence–enabled electrocardiogram,

Zachi I. Attia, Suraj Kapa, Francisco Lopez-Jimenez, Paul M. McKie, Dorothy J. Ladewig, Gaurav Satam, Patricia A. Pellikka, Maurice Enriquez-Sarano, Pe- ter A. Noseworthy, Thomas M. Munger, Samuel J. Asir- vatham, Christopher G. Scott, Rickey E. Carter, and Paul A. Friedman, “Screening for cardiac contrac- tile dysfunction using an artificial intelligence...

2019
[8]

EN- CASE: An ENsemble ClASsifiEr for ECG classification using expert features and deep neural networks,

Shenda Hong, Meng Wu, Yuxi Zhou, Qingyun Wang, Junyuan Shang, Hongyan Li, and Junqing Xie, “EN- CASE: An ENsemble ClASsifiEr for ECG classification using expert features and deep neural networks,” in2017 Computing in Cardiology (CinC), Sept. 2017, pp. 1–4, ISSN: 2325-887X

2017

Show all 22 references

[9]

MINA: Multilevel Knowledge-Guided Attention for Modeling Electrocardiography Signals,

Shenda Hong, Cao Xiao, Tengfei Ma, Hongyan Li, and Jimeng Sun, “MINA: Multilevel Knowledge-Guided Attention for Modeling Electrocardiography Signals,” inProceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence, Macao, China, Aug. 2019, pp. 588...

2019
[10]

Reading Your Heart: Learning ECG Words and Sentences via Pre-training ECG Language Model,

Jiarui Jin, Haoyu Wang, Hongyan Li, Jun Li, Jiahui Pan, and Shenda Hong, “Reading Your Heart: Learning ECG Words and Sentences via Pre-training ECG Language Model,” Feb. 2025, arXiv:2502.10707 [cs]

2025 arXiv
[11]

Synthetic ECG signals gener- ation: A scoping review,

Beatrice Zanchi, Giuliana Monachino, Luigi Fiorillo, Giulio Conte, Angelo Auricchio, Athina Tzovara, and Francesca D. Faraci, “Synthetic ECG signals gener- ation: A scoping review,”Computers in Biology and Medicine, vol. 184, pp. 109453, Jan. 2025

2025
[12]

Region-Disentangled Diffusion Model for High- Fidelity PPG-to-ECG Translation,

Debaditya Shome, Pritam Sarkar, and Ali Etemad, “Region-Disentangled Diffusion Model for High- Fidelity PPG-to-ECG Translation,”Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 13, pp. 15009–15019, Mar. 2024

2024
[13]

ECGTwin: Person- alized ECG Generation Using Controllable Diffusion Model,

Yongfan Lai, Bo Liu, Xinyan Guan, Qinghao Zhao, Hongyan Li, and Shenda Hong, “ECGTwin: Person- alized ECG Generation Using Controllable Diffusion Model,” Aug. 2025, arXiv:2508.02720 [cs]

2025
[14]

Generating Images with Multimodal Language Mod- els,

Jing Yu Koh, Daniel Fried, and Ruslan Salakhutdinov, “Generating Images with Multimodal Language Mod- els,” Oct. 2023, arXiv:2305.17216 [cs]

2023 arXiv
[15]

Making LLaMA SEE and Draw with SEED Tokenizer,

Yuying Ge, Sijie Zhao, Ziyun Zeng, Yixiao Ge, Chen Li, Xintao Wang, and Ying Shan, “Making LLaMA SEE and Draw with SEED Tokenizer,” Oct. 2023, arXiv:2310.01218 [cs]

2023 arXiv
[16]

WISE: A World Knowledge-Informed Semantic Evaluation for Text-to- Image Generation,

Yuwei Niu, Munan Ning, Mengren Zheng, Weiyang Jin, Bin Lin, Peng Jin, Jiaqi Liao, Chaoran Feng, Kun- peng Ning, Bin Zhu, and Li Yuan, “WISE: A World Knowledge-Informed Semantic Evaluation for Text-to- Image Generation,” May 2025, arXiv:2503.07265 [cs]

2025 arXiv
[17]

BLIP3-o: A Family of Fully Open Unified Mul- timodal Models-Architecture, Training and Dataset,

Jiuhai Chen, Zhiyang Xu, Xichen Pan, Yushi Hu, Can Qin, Tom Goldstein, Lifu Huang, Tianyi Zhou, Saining Xie, Silvio Savarese, Le Xue, Caiming Xiong, and Ran Xu, “BLIP3-o: A Family of Fully Open Unified Mul- timodal Models-Architecture, Training and Dataset,” May 2025, arXiv:25...

2025 arXiv
[18]

Teach Multimodal LLMs to Comprehend Electrocar- diographic Images,

Ruoqi Liu, Yuelin Bai, Xiang Yue, and Ping Zhang, “Teach Multimodal LLMs to Comprehend Electrocar- diographic Images,” Oct. 2024, arXiv:2410.19008 [eess]

2024 arXiv
[19]

ECG-Chat: A Large ECG-Language Model for Cardiac Disease Diagnosis,

Yubao Zhao, Jiaju Kang, Tian Zhang, Puyu Han, and Tong Chen, “ECG-Chat: A Large ECG-Language Model for Cardiac Disease Diagnosis,” Apr. 2025, arXiv:2408.08849 [eess]

2025 arXiv
[20]

GEM: Empowering MLLM for Grounded ECG Understanding with Time Series and Images,

Xiang Lan, Feng Wu, Kai He, Qinghao Zhao, Shenda Hong, and Mengling Feng, “GEM: Empowering MLLM for Grounded ECG Understanding with Time Series and Images,” Mar. 2025, arXiv:2503.06073 [cs]

2025
[21]

DiffuSETS: 12-Lead ECG generation conditioned on clinical text reports and patient-specific information,

Yongfan Lai, Jiabo Chen, Qinghao Zhao, Deyun Zhang, Yue Wang, Shijia Geng, Hongyan Li, and Shenda Hong, “DiffuSETS: 12-Lead ECG generation conditioned on clinical text reports and patient-specific information,” Patterns, p. 101291, June 2025

2025
[22]

MIMIC-IV-ECG: Diagnostic Electrocardiogram Matched Subset,

Brian Gow, Tom Pollard, Larry A Nathanson, Al- istair Johnson, Benjamin Moody, Chrystinne Fernan- des, Nathaniel Greenbaum, Jonathan W Waks, Paras- tou Eslami, Tanner Carbonati, Ashish Chaudhari, Eliza- beth Herbst, Dana Moukheiber, Seth Berkowitz, Roger Mark, and Steven Horng...

2023

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.

[1] [1]

INTRODUCTION The electrocardiogram (ECG) has become a crucial tool in the clinical diagnosis of cardiac diseases, owing to its non- invasive nature, low cost, and ease of acquisition [1, 2, 3]. While deep learning methods have shown promising results in tasks such as cardiac anomaly detection [4, 5, 6] and rare ECG synthesis [7, 8, 9], these methods are t...

arXiv 2025

[2] [2]

Generate an ECG of

METHODOLOGY 2.1. Learning to Understand ECG Given an ECG signals, an ECG imagex, and its textual captiony, our goal is to fully fine-tune a large language model (LLM) such that it can jointly reason over ECG data and textual inputs, thereby generating clinically meaning- ful responses grounded in ECG interpretation. To achieve this, we leverage the ECG-Gr...

[3] [3]

Datasets We primarily employ the ECG-Grounding and MIMIC-IV- ECG datasets for training and validation in UniECG under- standing and ECG generation tasks

EXPERIMENT AND ANALYSIS 3.1. Datasets We primarily employ the ECG-Grounding and MIMIC-IV- ECG datasets for training and validation in UniECG under- standing and ECG generation tasks. The descriptions of these two datasets are provided as follows: MIMIC-IV-ECG.This publicly available dataset [18] provides a comprehensive collection of 800,035 multiple- lea...

[4] [4]

CONCLUSION In this paper, we introduce UniECG, the first unified model for ECG that is capable of performing both evidence-based ECG understanding and text-guided ECG generation. We propose a two-stage training strategy that equips existing ECG MLLMs with both interpretive and generative capabil- ities, thereby effectively expanding the capability frontie...

[5] [5]

Opportunities and challenges of deep learning methods for electrocardiogram data: A system- atic review,

Shenda Hong, Yuxi Zhou, Junyuan Shang, Cao Xiao, and Jimeng Sun, “Opportunities and challenges of deep learning methods for electrocardiogram data: A system- atic review,”Computers in Biology and Medicine, vol. 122, pp. 103801, July 2020

2020

[6] [6]

Deep learning in ECG diagnosis: A review,

Xinwen Liu, Huan Wang, Zongjin Li, and Lang Qin, “Deep learning in ECG diagnosis: A review,” Knowledge-Based Systems, vol. 227, pp. 107187, Sept. 2021

2021

[7] [7]

Screening for cardiac contrac- tile dysfunction using an artificial intelligence–enabled electrocardiogram,

Zachi I. Attia, Suraj Kapa, Francisco Lopez-Jimenez, Paul M. McKie, Dorothy J. Ladewig, Gaurav Satam, Patricia A. Pellikka, Maurice Enriquez-Sarano, Pe- ter A. Noseworthy, Thomas M. Munger, Samuel J. Asir- vatham, Christopher G. Scott, Rickey E. Carter, and Paul A. Friedman, “Screening for cardiac contrac- tile dysfunction using an artificial intelligence...

2019

[8] [8]

EN- CASE: An ENsemble ClASsifiEr for ECG classification using expert features and deep neural networks,

Shenda Hong, Meng Wu, Yuxi Zhou, Qingyun Wang, Junyuan Shang, Hongyan Li, and Junqing Xie, “EN- CASE: An ENsemble ClASsifiEr for ECG classification using expert features and deep neural networks,” in2017 Computing in Cardiology (CinC), Sept. 2017, pp. 1–4, ISSN: 2325-887X

2017